Re: [rohc] Formal notation, combined thread

[Carsten Bormann <cabo@tzi.org>]

> Both approaches have their advantages - the EPIC-
> lite approach is always more efficient, but the self-
> describing values approach is simpler and less
> memory-intensive.  From the notation perspective
> we should therefore make sure that both approaches
> are supported, and leave it up to profile writers to
> select the correct approach for a given situation.
I envision it like this:

Given a set of rules like:

a --> b | c.
b --> d | e.
c --> f | g.
and so on,

you would wrap this whole thing into something like:

z --> epic(a).

Now this would collect the information in the notation tree opened by 
the above rules and prefix the values generated by a to g with an EPIC 
discriminator.  I.e., a discriminator value is a value still to be 
compressed, and you have to use a meta-encoding method (an encoding 
method that is supplied an encoding method as a parameter) to convert 
it into actual compressed values.  It would be an error to have 
discriminator values left at the root encoding method.

If you want to use a different discriminator encoding method, just say

z --> huffman(a).
or
z --> arith(a).

Does that work?

Gruesse, Carsten

PS.: The above "|" notation is missing the probabilities.  So it will 
look different in the end.

_______________________________________________
Rohc mailing list
Rohc@ietf.org
https://www1.ietf.org/mailman/listinfo/rohc